Blog Musk Touts ‘Quantum Leap” in Full Self-Driving Performance

A “quantum leap” improvement is coming to Tesla’s Autopilot software in six to 10 weeks, Chief Executive Elon Musk said a tweet.

Musk called the new software a “fundamental architectural rewrite, not an incremental tweak.”

https://twitter.com/x/status/1294374864657162240

Musk said his personal car is running a “bleeding edge alpha build” of the software, which he also mentioned during Tesla’s Q2 earnings. 

“So it’s almost getting to the point where I can go from my house to work with no interventions, despite going through construction and widely varying situations,” Musk said on the earnings call. “So this is why I am very confident about full self-driving functionality being complete by the end of this year, is because I’m literally driving it.”

Tesla’s Full Self-Driving software has been slow to roll out against the company’s promises. Musk previously said a Tesla would drive from Los Angeles to New York using the Full Self Driving feature by the end of 2019. The company didn’t meet that goal. So, it will be interesting to see the state of Autopilot at the end of 2020.

 

[mikes_fsd]

One more thing...

That was in post #264.
That was in post #311.

Did you actually expect someone to follow your train of thought/argument 47 posts separated (from #264 to #311) in the thread? One from August 25th the other September 5th...

That's a very "robust" argument you got there.

 

[DanCar]

how they will resolve conflicts.

If I were designing the system, I would have each system output a confidence score. If there are two systems then choose whichever has the highest confidence. If there are three systems, then it gets more complicated. For example do you choose the two that are most correlated?

 

[diplomat33]

One more thing...

Did you actually expect someone to follow your train of thought/argument 47 posts separated (from #264 to #311) in the thread? One from August 25th the other September 5th...

That's a very "robust" argument you got there.

I was merely providing the proper context because my quote which you say was me stating a fact was actually a clarification on a previous opinion. Another poster, SEXY cars, was specifically asking me about something I wrote in post #264.

Also, in post #311 which you quote, I reiterate that it is my opinion:

So I think they could do self-driving with just cameras.

 

[mikes_fsd]

Also, in post #311 which you quote, I reiterate that it is my opinion:

There is only one post I am quoting that is #311.
Just because you found a convenient scapegoat to point at does not change the fact of what you wrote in post #311.

 

[diplomat33]

There is only one post I am quoting that is #311.
Just because you found a convenient scapegoat to point at does not change the fact of what you wrote in post #311.
View attachment 585487

And what is wrong with my post #311?

"Both. Waymo can self-drive if just one lidar fails or if all lidar fails" is not a statement of fact. You are taking it out of context. It was in response to SEXY Cars's question regarding my opinion in post #264. He asked if I think Waymo can self-drive if 1 lidar fails or if all lidar fails. I was clarifying my earlier opinion.

"Waymo has camera vision that can identify objects..." is a valid fact.

"So I think they could do self-driving with just cameras" is an opinion since I say "I think". It is an opinion based on the evidence that I just laid out, ie that Waymo has camera vision that can detect objects etc...

"Our latest long range cameras..." is a quote directly from Waymo which is part of my evidence to support my opinion.

I provided my opinion and some of my evidence to support it.

 

[powertoold]

Some of the people here underestimate what Tesla has achieved with AP thus far.

 

[svusa]

I think the fight is useless, we all have opinions even about how the universe functions!!! IMO: Bottomline is that L4 Tesla autopilot 5 years later shouldn't make anyone happy who paid for FSD. Those who bought HW 2.0 cars, the value of your car has already been depreciated and it will be considered old by any standards. Whats point of having L4 when you are already due for a refresh!

 

[Cheburashka]

Elon Musk Says Tesla Cars Will Reach 620 Miles On A Single Charge “Within A Year Or Two,” Be Fully Autonomous In “Three Years”

 

[powertoold]

Elon Musk Says Tesla Cars Will Reach 620 Miles On A Single Charge “Within A Year Or Two,” Be Fully Autonomous In “Three Years”

He's technically right. NoA is "fully autonomous" on highways, and people have driven Teslas to 1000km using hypermiling. Lol

 

[diplomat33]

Elon Musk Says Tesla Cars Will Reach 620 Miles On A Single Charge “Within A Year Or Two,” Be Fully Autonomous In “Three Years”

A great example of how wrong Elon's predictions are. He made those predictions back in 2015.

 

[Cheburashka]

A great example of how wrong Elon's predictions are. He made those predictions back in 2015.

Yeah well, it's 2020 and we got a green light chime.

FSD must be around the corner! Quantum leap incoming.

 

[mikes_fsd]

A great example of how wrong Elon's predictions are. He made those predictions back in 2015.

Yes, but I first believed Google's / Waymo's predictions... Google's Sergey Brin: You'll ride in robot cars within 5 years
In case you missed it... the article is from 2012!

 

[mikes_fsd]

This actually brings me to my next point.

In 2012 - Google proudly announced to the world.

Our vehicles, of which about a dozen are on the road at any given time, have now completed more than 300,000 miles of testing. They've covered a wide range of traffic conditions, and there hasn't been a single accident under computer control.

Google self-driving car 300,000 miles -- and counting

In 2012 - Tesla was not even offering Autopilot 1 (MobilEye solution) on its vehicles.

The stark contrast of progress in the time since then is quite eye-opening!

 

[Daniel in SD]

The stark contrast of progress in the time since then is quite eye-opening!

In 2012 Google's vehicles were way more capable than the demo from Autonomy Day (which itself showed zero progress from the 2016 demo). It turns out that making a car that can self drive 100 miles between failures is easy compared to making one that is safe enough to operate without human supervision. I suspect Tesla will also discover this.

 

[mikes_fsd]

In 2012 Google's vehicles were way more capable than the demo from Autonomy Day (which itself showed zero progress from the 2016 demo). It turns out that making a car that can self drive 100 miles between failures is easy compared to making one that is safe enough to operate without human supervision. I suspect Tesla will also discover this.

I think Tesla has discovered this already. I also believe that is the reason that a fundamental rewrite of their AP code was needed.

We will find out soon enough.

 

[Cheburashka]

In 2012 Google's vehicles were way more capable than the demo from Autonomy Day (which itself showed zero progress from the 2016 demo). It turns out that making a car that can self drive 100 miles between failures is easy compared to making one that is safe enough to operate without human supervision. I suspect Tesla will also discover this.

Yeah I was going to say.... haven't really seen much progress beyond promises. We only got speed sign recognition last week. We've had incremental and slow progress for the past few years but they haven't been able to solve problems such as phantom braking which isn't giving me too much confidence. I hope I'm wrong.

 

[mikes_fsd]

I hope I'm wrong.

I hope so too!

 

[DanCar]

In 2012 Google's vehicles were way more capable than ...

How do we know this?

 

[diplomat33]

Yes, but I first believed Google's / Waymo's predictions... Google's Sergey Brin: You'll ride in robot cars within 5 years
In case you missed it... the article is from 2012!

This actually brings me to my next point.

In 2012 - Google proudly announced to the world.

Google self-driving car 300,000 miles -- and counting

In 2012 - Tesla was not even offering Autopilot 1 (MobilEye solution) on its vehicles.

The stark contrast of progress in the time since then is quite eye-opening!

And Sergey Brin was actually correct in his prediction. Waymo got their first robotaxi in 2015. And in 5 years, Waymo did achieve robot cars that the public can ride in. Yes, it has taken them longer to deploy robotaxis wide because of how difficult reliable L4/L5 autonomy is but Waymo did achieve robotaxis that the public can use when they said they would.

It shows how far ahead Waymo is. When Tesla was still on AP1, Google had FSD prototypes. And when Tesla was announcing AP2, Waymo already had robotaxis on public roads! Tesla still has no FSD and Waymo has L4 robotaxis that can operate with no human driver in the car at all.

In the time it has taken Tesla to go from AP1 to AP3 and still not achieve any FSD yet, Waymo achieved L4 robotaxis.

 

[diplomat33]

How do we know this?

Because Google had self-driving prototypes that could drive on highways and city streets, make turns at intersections etc...

This is from 2014 with Google's 2nd Generation FSD car:

By the way, watch the video and see what the Google FSD car could do when Tesla was just releasing AP1.